agent-inspect 6.28.0 → 6.29.0

package/CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changelog
 
+## 6.29.0
+
+### Minor Changes
+
+- 0e5732e: Provider usage fidelity: preserve cache-read (`cached`), cache-write (`cacheWrite`), and reasoning token fields across core normalization and adapters. Add isolated AI SDK peer matrix (`ai@^6` required; AI SDK 7 soft-blocked pending telemetry API migration). No pricing engine; root Node remains `>=20`.
+
 ## 6.28.0
 
 ### Minor Changes

package/README.md

@@ -212,7 +212,7 @@ The root package is enough for custom capture, the CLI, checks, and Evidence wor
 
 ## Status and documentation
 
-**Current published baseline:** **6.28.0** · persisted schema `1.0` · Node.js `>=20` · MIT.
+**Current published baseline:** **6.29.0** · persisted schema `1.0` · Node.js `>=20` · MIT.
 
 Legacy v0.1 and v0.2 traces remain readable. Check the npm badge and [changelog](CHANGELOG.md) for the current published version.
 

package/docs/LIMITATIONS.md

@@ -81,7 +81,7 @@ This document states what AgentInspect **does not** provide today. It complement
 
 ## Economics
 
-- **No cost engine**: no pricing tables, invoice-grade usage, or provider billing reconciliation.
+- **No cost engine**: no pricing tables, invoice-grade usage, or provider billing reconciliation. Usage fidelity may preserve cache-read (`cached`), cache-write (`cacheWrite`), and reasoning token counts when providers report them; those fields are observational only.
 - **Token usage is supplied metadata only**: AgentInspect may display `input`, `output`, `total`, and `cached` counts when callers/adapters provide them; core does not count tokens or infer provider billing.
 
 ## Local observability commands (v1.4.0)

package/docs/SCHEMA.md

@@ -264,7 +264,7 @@ Canonical samples: `fixtures/traces-v0.2/*.jsonl` (validated by `pnpm fixtures:c
 | `attributes` | no | Shallow metadata bag (redaction-ready) |
 | `inputSummary` / `outputSummary` | no | Truncated previews when explicitly captured |
 | `error` | no | `{ name?, message, code? }` when `status: "error"` |
-| `tokenUsage` | no | `{ input?, output?, total?, cached? }` when supplied/known |
+| `tokenUsage` | no | `{ input?, output?, total?, cached?, cacheWrite?, reasoning? }` when supplied/known. `cached` is cache-read; `cacheWrite` / `reasoning` are informational and are never added to `total`. No pricing fields. |
 | `trace` | no | Optional `{ traceId?, spanId?, parentSpanId? }` for future OTel alignment |
 
 Programmatic helpers: see [API.md](./API.md) §13 (experimental persisted-event foundation).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-inspect",
-  "version": "6.28.0",
+  "version": "6.29.0",
   "license": "MIT",
   "type": "module",
   "description": "Local evidence debugger and trajectory-test toolkit for TypeScript AI agents — execution trees, TraceContract checks, Evidence v2, and read-only MCP",
@@ -250,6 +250,7 @@
     "demo:generate": "node scripts/demo-generate.mjs",
     "demo:verify": "node scripts/demo-verify.mjs",
     "compat:smoke": "node scripts/compat-smoke.mjs",
+    "compat:ai-sdk-matrix": "node scripts/ai-sdk-peer-matrix.mjs",
     "fixtures:check": "node scripts/validate-fixtures.mjs",
     "recipes:check": "node scripts/validate-recipes.mjs",
     "perf:baseline": "node scripts/performance-baseline.mjs",

package/packages/cli/dist/{chunk-JVQ4LTHR.mjs → chunk-D7PT6TXM.mjs}

@@ -94,6 +94,8 @@ function isPersistedTokenUsage(value) {
   if (!isOptionalNonNegativeNumber(value.output)) return false;
   if (!isOptionalNonNegativeNumber(value.total)) return false;
   if (!isOptionalNonNegativeNumber(value.cached)) return false;
+  if (!isOptionalNonNegativeNumber(value.cacheWrite)) return false;
+  if (!isOptionalNonNegativeNumber(value.reasoning)) return false;
   return true;
 }
 function isPersistedTraceContext(value) {
@@ -1796,7 +1798,13 @@ function buildRunSummary(events) {
         tokensInput: isNonNegativeFiniteNumber(s.metadata?.tokens?.input) ? s.metadata.tokens.input : void 0,
         tokensOutput: isNonNegativeFiniteNumber(s.metadata?.tokens?.output) ? s.metadata.tokens.output : void 0,
         tokensTotal: isNonNegativeFiniteNumber(s.metadata?.tokens?.total) ? s.metadata.tokens.total : void 0,
-        tokensCached: isNonNegativeFiniteNumber(s.metadata?.tokens?.cached) ? s.metadata.tokens.cached : void 0
+        tokensCached: isNonNegativeFiniteNumber(s.metadata?.tokens?.cached) ? s.metadata.tokens.cached : void 0,
+        tokensCacheWrite: isNonNegativeFiniteNumber(
+          s.metadata?.tokens?.cacheWrite
+        ) ? s.metadata.tokens.cacheWrite : void 0,
+        tokensReasoning: isNonNegativeFiniteNumber(
+          s.metadata?.tokens?.reasoning
+        ) ? s.metadata.tokens.reasoning : void 0
       });
     }
   }
@@ -1820,9 +1828,13 @@ function buildRunSummary(events) {
   let totalTokensOutput = 0;
   let totalTokensTotal = 0;
   let totalTokensCached = 0;
+  let totalTokensCacheWrite = 0;
+  let totalTokensReasoning = 0;
   let tokenBearingSteps = 0;
   let stepsWithKnownTotal = 0;
   let hasCachedTokens = false;
+  let hasCacheWriteTokens = false;
+  let hasReasoningTokens = false;
   const depthCache = /* @__PURE__ */ new Map();
   for (const [id, s] of steps.entries()) {
     totalSteps += 1;
@@ -1837,7 +1849,7 @@ function buildRunSummary(events) {
         longestStep = { name: s.name, durationMs: s.durationMs, type: s.type };
       }
     }
-    if (s.tokensInput !== void 0 || s.tokensOutput !== void 0 || s.tokensTotal !== void 0 || s.tokensCached !== void 0) {
+    if (s.tokensInput !== void 0 || s.tokensOutput !== void 0 || s.tokensTotal !== void 0 || s.tokensCached !== void 0 || s.tokensCacheWrite !== void 0 || s.tokensReasoning !== void 0) {
       tokenBearingSteps += 1;
       if (s.tokensInput !== void 0) totalTokensInput += s.tokensInput;
       if (s.tokensOutput !== void 0) totalTokensOutput += s.tokensOutput;
@@ -1852,6 +1864,14 @@ function buildRunSummary(events) {
         totalTokensCached += s.tokensCached;
         hasCachedTokens = true;
       }
+      if (s.tokensCacheWrite !== void 0) {
+        totalTokensCacheWrite += s.tokensCacheWrite;
+        hasCacheWriteTokens = true;
+      }
+      if (s.tokensReasoning !== void 0) {
+        totalTokensReasoning += s.tokensReasoning;
+        hasReasoningTokens = true;
+      }
     }
   }
   const summary = {
@@ -1871,7 +1891,9 @@ function buildRunSummary(events) {
         input: totalTokensInput,
         output: totalTokensOutput,
         ...stepsWithKnownTotal === tokenBearingSteps ? { total: totalTokensTotal } : {},
-        ...hasCachedTokens ? { cached: totalTokensCached } : {}
+        ...hasCachedTokens ? { cached: totalTokensCached } : {},
+        ...hasCacheWriteTokens ? { cacheWrite: totalTokensCacheWrite } : {},
+        ...hasReasoningTokens ? { reasoning: totalTokensReasoning } : {}
       }
     } : {}
   };
@@ -2384,6 +2406,12 @@ function renderRunWhat(summary, options = {}) {
     if (summary.totalTokens.cached !== void 0) {
       tokenParts.push(`${summary.totalTokens.cached} cached`);
     }
+    if (summary.totalTokens.cacheWrite !== void 0) {
+      tokenParts.push(`${summary.totalTokens.cacheWrite} cache-write`);
+    }
+    if (summary.totalTokens.reasoning !== void 0) {
+      tokenParts.push(`${summary.totalTokens.reasoning} reasoning`);
+    }
     lines.push(`Tokens: ${tokenParts.join(" / ")}`);
   }
   if (showCorrelation && summary.correlation) {
@@ -5186,6 +5214,9 @@ var SAFE_USAGE_LEAF_KEYS = /* @__PURE__ */ new Set([
   "output",
   "total",
   "cached",
+  "cachewrite",
+  "cache_write",
+  "reasoning",
   "input_tokens",
   "inputtokens",
   "output_tokens",
@@ -5195,7 +5226,13 @@ var SAFE_USAGE_LEAF_KEYS = /* @__PURE__ */ new Set([
   "prompt_tokens",
   "prompttokens",
   "completion_tokens",
-  "completiontokens"
+  "completiontokens",
+  "cache_read",
+  "cacheread",
+  "cache_creation",
+  "cachecreation",
+  "reasoning_tokens",
+  "reasoningtokens"
 ]);
 var DEFAULT_SECRET_PATTERNS = [
   { id: "bearer-token", pattern: /Bearer\s+[A-Za-z0-9._~+/-]{12,}=*/ },
@@ -8792,16 +8829,20 @@ function normalizeTokenUsage(value) {
   const output = nonNegativeFinite(value.output);
   const suppliedTotal = nonNegativeFinite(value.total);
   const cached = nonNegativeFinite(value.cached);
+  const cacheWrite = nonNegativeFinite(value.cacheWrite);
+  const reasoning = nonNegativeFinite(value.reasoning);
   const derivedTotal = input !== void 0 && output !== void 0 && Number.isFinite(input + output) ? input + output : void 0;
   const total = suppliedTotal ?? derivedTotal;
-  if (input === void 0 && output === void 0 && total === void 0 && cached === void 0) {
+  if (input === void 0 && output === void 0 && total === void 0 && cached === void 0 && cacheWrite === void 0 && reasoning === void 0) {
     return void 0;
   }
   return {
     ...input !== void 0 ? { input } : {},
     ...output !== void 0 ? { output } : {},
     ...total !== void 0 ? { total } : {},
-    ...cached !== void 0 ? { cached } : {}
+    ...cached !== void 0 ? { cached } : {},
+    ...cacheWrite !== void 0 ? { cacheWrite } : {},
+    ...reasoning !== void 0 ? { reasoning } : {}
   };
 }
 
@@ -9960,6 +10001,8 @@ function readOpenInferenceTokenUsage(attributes) {
   const completion = attributes["llm.token_count.completion"];
   const total = attributes["llm.token_count.total"];
   const cached = attributes["llm.token_count.prompt_details.cache_read"];
+  const cacheWrite = attributes["llm.token_count.prompt_details.cache_write"] ?? attributes["llm.token_count.prompt_details.cache_creation"];
+  const reasoning = attributes["llm.token_count.completion_details.reasoning"] ?? attributes["llm.token_count.completion_details.reasoning_tokens"];
   const usage = {};
   if (typeof prompt === "number" && Number.isFinite(prompt) && prompt >= 0) {
     usage.input = prompt;
@@ -9973,6 +10016,12 @@ function readOpenInferenceTokenUsage(attributes) {
   if (typeof cached === "number" && Number.isFinite(cached) && cached >= 0) {
     usage.cached = cached;
   }
+  if (typeof cacheWrite === "number" && Number.isFinite(cacheWrite) && cacheWrite >= 0) {
+    usage.cacheWrite = cacheWrite;
+  }
+  if (typeof reasoning === "number" && Number.isFinite(reasoning) && reasoning >= 0) {
+    usage.reasoning = reasoning;
+  }
   if (usage.total === void 0 && usage.input !== void 0 && usage.output !== void 0) {
     usage.total = usage.input + usage.output;
   }
@@ -12337,5 +12386,5 @@ function renderGateReport(result, options = {}) {
 }
 
 export { COHORT_METRIC_IDS, DEFAULT_SUITE_ARTIFACTS_DIR, EVIDENCE_FORMAT_VERSION, EVIDENCE_HTML_FILENAME, EVIDENCE_MANIFEST_FILENAME, Redactor, TraceDirectory, TraceReadError, TreeBuilder, aggregateBundleSafeStatus, aggregateSessionCheckResults, analyzeCohort, applyProfileMetadataCaps, assertBundlePathContained, assertEvidenceRelativePath, buildActivitySummary, buildBundleMetadata, buildBundleSummaryMarkdown, buildEvidenceCausalFailureViewHtml, buildEvidenceCiPackage, buildEvidenceCircuitViewHtml, buildEvidenceContractsViewHtml, buildEvidenceDiffViewHtml, buildEvidenceHtmlShell, buildEvidenceManifest, buildEvidenceOutcomesViewHtml, buildEvidenceProvenanceViewHtml, buildEvidenceSafetyViewHtml, buildEvidenceTimelineViewHtml, buildEvidenceToolsLlmViewHtml, buildEvidenceTreeViewHtml, buildLocalExplanation, buildPlaceholderArtifact, buildRunSummary, buildRunTimeline, buildRunWhatSummary, buildSessionIndex, buildTraceStats, buildZipArchive, bundleFailsOnSafety, bundleRunAssetRelativePath, collectTraceSchemaVersions, compactAttributes, createBaselineRegressionRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRunDepthRule, createRunDurationRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolUsageRule, defaultBundleOutputPath, defaultSuiteConfigTemplate, diffRuns, diffTraceEvents, enrichSessionRunRecord, escapeHtml, escapeMarkdown, extractMetadata, extractOutcomesFromTraceEvents, filterMetasBySessionScope, filterTraces, flattenTree, formatDuration2 as formatDuration, formatStepLabel, formatTimestamp, gateHasThresholds, getIndent, getTraceFilePath, inferEvidenceFileRole, isAgentInspectTrace, isPersistedInspectEvent, loadSessionRunRecords, loadSuiteConfig, loadTraceMetadataList, manualTraceEventsToComparableRun, nanoid, normalizeBundleOutputPath, openTrace, parseCohortMetricList, parseDuration, parseDurationFilter, parseGateList, parseTraceJsonl, persistedInspectEventsToTraceEvents, renderActivitySummaryHuman, renderCohortReport, renderErrorLine, renderGateReport, renderObservedOutcomesHtml, renderObservedOutcomesMarkdown, renderRunDiff, renderRunWhat, renderStepLine, renderSuiteReport, renderTimeline, renderTraceStats, resolveBundleRunIds, resolveRedactionProfile, resolveSuiteTemplate, resolveTraceDir, runGate, runSuite, runTraceChecks, safeString, sanitizeBundleRunId, searchTraces, serializeEvidenceManifest, sha256Hex, stableJson, summarizeObservedOutcomes, summarizeSemanticParity, traceEventToPersistedInspectEvent, truncateName, truncateStringForProfile, validateEvent, validateSuiteConfig, verifyEvidenceDirectory, zeroKinds };
-//# sourceMappingURL=chunk-JVQ4LTHR.mjs.map
-//# sourceMappingURL=chunk-JVQ4LTHR.mjs.map
+//# sourceMappingURL=chunk-D7PT6TXM.mjs.map
+//# sourceMappingURL=chunk-D7PT6TXM.mjs.map